Prognostic factors (PF) in advanced hepatocellular carcinoma (AHCC): Multivariate analysis and comparison between staging systems (SS) in patients (pts) treated at Memorial Sloan-Kettering Cancer Center (MSKCC)

Abstract

4601 Background: Several SS have been proposed in hepatocellular carcinoma. These include TNM, Okuda, Cancer of the Liver Italian Program (CLIP), Chinese University Prognostic Index (CUPI), and Barcelona Clinic Liver Cancer (BCLC). There is no consensus as to what constitutes the best SS for use by oncologists for pts with AHCC with no locoregional therapy options. We propose to define the PF and compare SS in this patient population. SS may help select pts for systemic therapy, predict outcome, and help in clinical trial design for AHCC. Methods: We retrospectively identified pts with AHCC treated at MSKCC between 2001 and 2006. Clinical, laboratory, tumor characteristics and all four SS were recorded. Survival (S) was measured from the date of development of AHCC to the date of death. S was estimated using Kaplan-Meier’s method, differences in S were tested using the log rank test. A Cox regression model was used for the multivariate analysis. A second Cox regression was done to compare SS and was expressed using the Akaike information (AI) criterion. AI helps determine which SS is the most informative of S. A low AI is favorable. Results: We identified 280 pts. Data on the first 101 pts analyzed are presented. Median age 61 years; 71% males, 29% females; 60% Caucasians, 9% Black, 24% Asians and 5% Hispanics. Etiologies included HCV 24%, HBV 38%, and alcohol 22%. Child Pugh score: A in 65% and B in 29% of pts. Multivariate analysis independent PF for S were albumin (p=0.0358), alkaline phosphatase (ALP) (p=0.001), identified etiology (p=0.008), abdominal pain (p=0.001) and liver tumor extent (more or less than 50% of the liver) (p=0.0043). AI ranked SS as follows: TNM 6th (588.991), TNM 5th (591.373), BCLC (541.095), Okuda (540.490), CLIP (537.8), and CUPI (526.483). CUPI S was 19.47 months (m) for low, 5.89 m for medium, and 1.36 m for high risk pts. Conclusions: Pts with AHCC who are treated by oncologists in this US-based population have distinct PF. CUPI provided the best prognostic information for our patient population. CUPI may be suggested as the SS to use clinically for AHCC. These results need prospective validation. No significant financial relationships to disclose.